DEV Community

T A Srinivas
T A Srinivas

Posted on

# 🛡️ Building an Enterprise AML/CFT Risk Intelligence Platform with Gemma 4

#ai #hacktoberfest #gemma #python

🛡️ Building an Enterprise AML/CFT Risk Intelligence Platform with Gemma 4

🏆 Hackathon Prize Track Alignment

This project explicitly targets the Best Open-Source AI Project Track and the Best Use of Gemma 4 Track. It builds a highly disciplined compliance architecture around the Reserve Bank of India (RBI) Master KYC Directions, using open-weight architectures to ensure 100% data residency and secure financial audit pipelines.

📊 Core Workspace Architecture

  • Stateless Semantic Reasoning Core: Integrates Google's Gemma 4 (gemma-4-26b-a4b-it) model combined with LangChain Vector RAG stores (FAISS) to parse legal regulation chunks dynamically.
  • Deterministic Compliance Math Engine: Employs parallelized Pandas ledger tracking to catch temporal cross-border spikes, layering velocity anomalies within 48-hour windows, and low-value cyber-fraud mule smurfing networks down to ₹1,000.
  • Zero-Trust IAM Security Matrix: Implements full multi-tab Role-Based Access Control (RBAC) in Streamlit, completely write-locking and hiding administrative screens from unauthorized roles.
  • Headless Vector PDF Reporting: Compiles formal compliance alerts and audit-ready forensic legal documents using a cloud-safe ReportLab binary generation matrix.

💻 Open-Source Source Code

Take a look at the complete decoupled system files, setup guidelines, and Mermaid deployment diagrams:

Top comments (0)